Mailman 3 March 2002 - Twisted

[Twisted-Python] This seems too good to be true... :-)
by Nils R Grotnes 18 Mar '02

18 Mar '02

Hi! I know I probably should lurk-a-lot-first, so please just tell me to shut up if I'm somehow out of line here... I like to state up front that I'm fresh, fresh and fresh. I switched to Debian just recently, I have just read my first Python book, and I didn't know this place existed until a week ago. The latter is what I wonder about in the Subject. How come it was so hard to find you? I have lurked quite a lot on comp.lang.python and c.l.p.announce, and I have surfed around on a lot of Python websites, but the first I ever heard of twisted was in c.l.p.a on the 9th. I'm very impressed with your setup and what I have seen of the code so far, so I figure _something_ must be wrong. :-) Ah, well. I hope I'm wrong, because I'd love both to use this stuff, and help out. I have coded a few (professional) projects before, but only in VB, so I'll keep my defiled hands off the code here until I've de-learned most of my bad habits... :-) Hopefully I can do some bug-hunting, testing and maybe even a little documentation after a while? (I dual boot to Win95 and have installed ActivePython there, if that can be useful.) My first try to do more than babble, will be a suggestion I couldn't find anything about in the archive; On my Debian, the package is installed in /usr/lib/python2.1/site-packages/twisted, but there's no twisted.pth file in the site-packages directory. After I added such a file (only containing the word "twisted" without the apostrophes) I now have twisted in sys.path when I start a Python shell. If it was missing for some good reason, I'd like to know it... Nils

2 3

[Twisted-Python] "3D Game Using Twisted?", or: "Ach du lieber..."
by Derek Simkowiak 15 Mar '02

15 Mar '02

Hello. My name is Derek Simkowiak, I've just been introduced to Twisted by Sean Riley (sean at twistedmatrix dot com). I'm using (and now contributing to) his PyUI widget set for my 3D game engine. When I told him I still needed to write the networking code, he pointed me here. I wanted to introduce myself, ask for some general guidance about a network protocol, and point out that the documentation titled "The Basics" at http://twistedmatrix.com/page.epy/docs/basics.html results in "no such page", which really sucks for somebody just getting started (like me). First, a little bit about my game engine, because it's the type of thing you twisted Twisted people might be interested in: o OpenGL for rendering o SDL for event management o Quake2 BSP map rendering (many more model formats to come soon) - includes almost all of the Id BSP culling tricks - includes fully-functional collision detection using the Q2 "bloated planes" algorithm, as figured out by Kekoa Proudfoot o Complete in-game GUI over the top of the 3D world, including transulency, using the PyUI widget set. Screenshot at: http://www.realloc.net/~dereks/targa_pyui1.jpg Also see the older shots at: http://www.realloc.net/~dereks/commputers/my_code/targa/ Also, the big piece that you may like: I've recently refactored the whole damn thing, and with a little help from SWIG my "engine" is now nothing more than some Python C extensions. My "main event loop" is just a simple Python script, and because all the [bottlenecking] rendering code is still C, I get the ~70 FPS I was getting when the entire engine was pure C. But now I have Python at my disposal, mua ha ha. (Also, note that all the libraries are cross-platform. I'm developing under Linux, but eventually this thing should run under MS-Windows, Mac OS X, AmigaOS, and even PlayStation2.) The big pieces I have left are the enemy A.I. and the networking code. This is a client-server engine. The light at the end of the tunnel is load-balanced Linux cluster (using LVS) serving up a persistent, 3D world (ala' PlanetSide), and maybe even me making some $$$ from a few hundred subscribers paying $5/month. Now you see where Twisted fits in -- I'm basically shooting for twisted.reality in real-time 3D. My concern is regarding bandwidth. My game protocol will be similar to the Quake2 protocol, meaning huge lists of 'entities' are updated with position and orientation data as often as possible. Such generic position updates will be done with UDP, and 'critical' packets (such as damage, explosions, jumps, messages, etc.) will be done with TCP. In summary, every single byte of network overhead matters -- nay, every BIT matters (npi). So I've browsed through the API docs, but I don't have any context to know what I'm looking at. I'll be downloading and installing the source soon, but in the meantime... any advice? Is there any particular package I should be looking at (other than twisted.internet)? Does Twisted's networking code cater to somebody who needs to, for example, do bitshifting and compression to absolutely minimize the bandwidth requirements? What approach would you take? Note that, using PyUI for the GUI and [possibly] Twisted for the server, I want to have things like in-game "IRC/ICQ", in-game "emails", and also in-game "voice chat" ala CounterStrike 1.3. Other wild ideas include in-game video conferencing, and dynamically-downloaded bits of Python code that would act as all-new, not-previously-imagined entities. (Note that by "in-game" I mean the 3D world is rendered, and the email/IRC/video client application is rendered on top of the 3D world with some translucency. The idea is to keep the player immersed.) Any advice is greatly appreciated. If anybody else is using Twisted for something similar, I'd love to hear about it. Thanks, Derek Simkowiak dereks at realloc dot net

9 19

[Twisted-Python] Issues with roots.Collections in regards to Coil
by Itamar Shtull-Trauring 15 Mar '02

15 Mar '02

(If this doesn't make sense, you can probably ignore it. At some point we might make a twisted-dev list.) Issues: 1) The problem where you click on a virtual host's sub-Resource in web coil and you can't configure it. Dunno why. (I fixed some similar effect bugs which were caused by *someone* using MOVED_PERMANENTLY and ignoring the fact this gets cached by browsers, but that's not the issue here.) 2) The UI assumes entities in a collection are mutable, i.e. first created, and then separately configured via a Configurator. This conflicts with stuff like making a Collection for configuring a mapping of strings (e.g. the mime types mapping in static.File). Possible solution is a custom Collection class for these situations that gets its own web coil UI. 3) Collections of Configurators or of configurables? I would think that they should stay the way they are -- collections of configurables. It's easy to wrap an object coming from a collection - that's what is done now. And it's easier to code the Collection this way, esp. since the configurable classes can be collections too, in cases where it makes sense (e.g. web.resource.Resource.) In most situations of course the Collections would only be coded in the coil plugin, but it's useful to be able to have the configurable class be a Collection. What were your issues, glyph?

2 2

[Twisted-Python] HTML in documentation
by Chris Armstrong 13 Mar '02

13 Mar '02

It has become apparent to me that we need to standardize the tags we use in HTML documentation for Twisted. This will make it easier to apply stylesheets to it, as well as use processing tools to convert it to other formats. Here is my proposal (This is especially aimed at Moshe, since he is the HTML purist of the bunch, and I want to make sure he's happy with these so he won't decide to rebel and start his own repository of docs again ;-)). So I would like it if at least Moshe and Glyph were to both reply to me in some fashion, even if with a simple "yes" or "no" (if it's a "no", a short explanation would be nice ;-)). Of course others are very much encouraged to chime in. The three of us aren't the only important ones around here, by a long shot. ;-) 1.) For code-snippets that are separated from paragraph content, we should either use the <code> tag, along with <pre>. for example: ---- <p> For example, this is how one defines a Resource: </p> <p> <code><pre> from twisted.web import resource class MyResource(resource.Resource): def render(self, request): return "Hello, world!" </pre></code> </p> ---- The <pre> shall not be assigned any special semantic value in processing tools or StyleSheets. It is only there so browsers will render it correctly. Special semantics shall only be assigned to <code>. 2.) For single-line code-snippets and attribute/method/module-names interpolated in a paragraph, <code> shall also be used, but without the <pre> tag. Also, module and class names should be specified with their fully-qualified name, and have a link to the API documentation, which is assumed to be in ../api/current/ (this is how it will be set up on the web-site). Regular attribute names don't need to be fully qualified (as this wouldn't make sense), but should still be enclosed in <code> tags. For example: ---- <p>To add a <a href="../api/current/twisted/web/widgets_Widget.html"><code>twisted.web.widgets.Widget</code></a> instance to a <a href="../api/current/twisted/web/widgets_Gadget.html"><code>twisted.web.widgets.Gadget</code></a> instance, do <code>myGadget.putWidget("widgetPath", MyWidget())</code>. </p> <p> (implementation note: the widgets are stored in the <code>gadgetInstance.widgets</code> attribute, which is a list.) </p> ---- I realize that adding the links to the API documentation everywhere is ugly, so I suppose it would be fine to only add the link to the first instance a module/classname shows up in the document. Just make sure there's a link in there somewhere. Maybe I'll hack on generate.py to automatically insert the links when it sees a fully-qualified twisted.* name... (but this doesn't mean one should completely disregard linking to API docs -- the idea is to have the HTML documents be browsable both with *or* without pre-processing) 3.) It goes without mentioning that one should use <hN> in a sane way to organize the document: the only <h1> shall be the title of the document, <h2> shall be sections, <h3> sub-sections, etc. 4.) Tags that have an optional closing tag should have the optional closing tag (ie, <li> and </li>). This makes it easier for some editors and processing tools to work with the code. This point isn't all that important. If it bothers someone enough, I can deal without having them. :-) 5.) Glyph says we should use XHTML, but he's a dork. Discuss. ;-) That's about it. Any other suggestions? If everyone agrees to this, I'll go through all the existing docs and update them with these rules (I doubt there will be many updates), and probably create a style-sheet to render them nicely. p.s. -- The only rules I notice that differ with Moshe's style in his existing documentation are 1.) he doesn't use <code> around multi-line code snippets and 2.) he doesn't have ending </li> tags. Is this a problem for you, Moshe?

1 1

[Twisted-Python] Automatic reconnection
by Andrew Bennetts 12 Mar '02

12 Mar '02

Hi all, I need a tcp.Client that can automatically reconnect when disconnected. I've written the following class, and it seems to work, but... * It doesn't handle unix sockets (that's ok for me, I'm using Windows, so I never had them in the first place) * I'm not sure what the "connector" is, and thus I suspect it won't for Clients that use it. * It's only lightly tested Is there a better way to do this (i.e. is there a way to do this already in Twisted, and I just haven't found it yet)? Otherwise, I'd be interested in knowing if anyone else would find this useful, or if maybe it should be included in Twisted. I suspect I'm not the only person who wants to be able to write a server which automatically tries to recover from "upstream" problems. -Andrew. Code: class ReconnectingClient(tcp.Client): """A tcp.Client that automatically reconnects when disconnected Overrideable attributes: * reconnectionDelay -- number of seconds to wait before reconnecting * maxReconnects -- maximum number of reconnects before giving up. The reconnection counter is reset after a successfull reconnection. Set this to None to keep retrying indefinitely. The protocol will have connectionLost called as normal when disconnected, but may get reconnected after that -- so if you need to reset anything in your protocol after a disconnection, do it in Protocol.connectionLost. """ reconnectionDelay = 3.0 maxReconnects = 5 def __init__(self, host, port, protocol, timeout=None, connector=None): tcp.Client.__init__(self, host, port, protocol, timeout, connector) self.reconnectAttempts = 0 self.reconnectProtocol = protocol def connectionLost(self): # Disconnect as normal tcp.Client.connectionLost(self) # Then schedule a reconnect if main.running: # FIXME: Is this test necessary? print 'Lost connection, reconnecting...' main.addTimeout(self.reconnect, self.reconnectionDelay) def reconnect(self): if not self.connected: if (self.maxReconnects is None or self.reconnectAttempts < self.maxReconnects): self.reconnectAttempts += 1 tcp.Connection.__init__(self, self.createInternetSocket(), self.reconnectProtocol) self.doConnect() main.addTimeout(self.reconnect, self.reconnectionDelay) else: print 'Reconnection failed after %d attempts. Giving up.' \ % (self.reconnectAttempts,) else: # Reset the reconnect counter if we're connected. self.reconnectAttempts = 0

2 2

[Twisted-Python] Re: Twisted protocol
by Andrew Bennetts 12 Mar '02

12 Mar '02

On Mon, Mar 11, 2002 at 06:51:57PM +0100, David wrote: > Hello spiv, this is supposed to be reaching you =) > I was told by radix on #python@opn that you had written "a little proxy > protocol which dispatches to multiple other protocol objects" in > twisted. This is exactly what i need for a project of mine. > I'd very much appriciate if you would share this little piece of code > with me or alternativly explain how it works. Heh. I'm CC'ing the list with this, just for the sake of it :) Here is the class. I created it for a case where I had a protocol that was sometimes inheriting from one Protocol, and sometimes another, and I didn't want to diddle with __bases__ or munge the two bases into one class (although perhaps that is the "better" solution?). If you like, imagine PacketBasedProtocolA/B to be basic.LineReceiver and Int32StringReceiver, where lineReceived and stringReceived are both aliased to packetReceived. class ProxiedPacketProtocol(Protocol): """General packet-based protocol. This class provides a simple way to use either PacketBasedProtocolA or PacketBasedProtocolB (or potentially others) as protocol for a data feed, without having to change which one you inherit from at runtime. It basically acts as a proxy protocol between the low-level protocol and your subclass. Simply override parentProtocol with the "parent" protocol (the default is PacketBasedProtocolA) and then when packetReceived is triggered in the parent, packetReceived is called in me. Inherit from this class, and override packetReceived.""" parentProtocol = PacketBasedProtocolA packets = {} # This is copied to the parent def __init__(self): # Create a private subclass of the parentProtocol so that we can # override its class-level data attribute(s). class ParentProtocol(self.parentProtocol): packets = self.packets # Create the parent protocol self.parent = ParentProtocol() # Raw data should get delivered to the parent... self.dataReceived = self.parent.dataReceived # ...and processed packets should come back to me self.parent.packetReceived = self.packetReceived self.parent.unknownPacket = self.unknownPacket # Data from the pinger needs to come via me self.parent.transport = self.transport def packetReceived(self, packet): """Override this. Called whenever the parentProtocol receives a packet.""" pass def unknownPacket(self, messageType, data): print "Received unknown packet of type '%s': %s" \ % (messageType, repr(data)) David, I hope this is what you were after :) It should be fairly easy to use and adapt -- but my question for the list is, is it worth the extra effort it would take to make this more general? i.e. take arbitrary lists of attributes to copy from the parent to the proxy (and vice versa?), instead of hard-coded names in __init__? My other question is, is this a useful enough concept to include it in Twisted? :) -Andrew.

2 1

[Twisted-Python] Win32 updates
by Itamar Shtull-Trauring 12 Mar '02

12 Mar '02

1) win32 process support is checked in and (lightly) tested. Just twisted.internet.win32.install() and it's there. For CGI support we're gonna need to write a small parser thing that can grok unix style "#!c:\python2.1\python.exe" interpreter lines. 2) Because the process tests install the win32 event loop, I discovered the win32 event loop breaks the the DNS client/server testing code, so currently these tests do *not* pass. Anyone know about issues with WSAEventSelect and UDP?

2 1

[Twisted-Python] docs
by Christopher Armstrong 12 Mar '02

12 Mar '02

I just shwacked all the HTML files in the 'doc/' directory and put them in the TwistedDocs repository. This repository will be populated with HTML documentation which have '.htc' suffixes -- these files *are* just plain old HTML. I copied/hacked 'generate.py' from the WebSite repository to be able to convert regular HTML to HTML-without-headers-and-footers, and also be able to extract the title from the <title> tags of the .htc document. the generate.py script generates "prettier" html documents with no suffix according to a template document which is passed to it on the command line. The generate script also converts all links to "foo.htc" to "foo", thus making linking transparent wether you're browsing un-processed docs or not. Anyway, you people who are simply writing the docs don't need to worry about anything -- just write html files with .htc suffixes and link between them as you would expect. :-) </babble> -- Chris Armstrong << radix(a)twistedmatrix.com >> http://twistedmatrix.com/users/carmstro.twistd/

1 1

[Twisted-Python] Yet another win32 process
by Itamar Shtull-Trauring 10 Mar '02

10 Mar '02

This one uses the win32 event loop for reading from stdout and stderr, and starts one thread per Process for writing to stdin. So we're down to one thread per Process instead of 3, thanks to Andrew's suggestion. """ Win32 process support. """ # Twisted imports from twisted.internet import abstract, main, win32 # System imports import win32api import win32pipe import win32file import win32process import win32security import win32con import win32event import pywintypes import msvcrt import os import sys import threading import Queue import string class Process(abstract.FileDescriptor): """A process that integrates with the Twisted event loop.""" buffer = '' def __init__(self, command, args, environment, path): # security attributes for pipes sAttrs = win32security.SECURITY_ATTRIBUTES() sAttrs.bInheritHandle = 1 # create pipes hStdin_r, self.hStdin_w = win32pipe.CreatePipe(sAttrs, 0) self.hStdout_r, hStdout_w = win32pipe.CreatePipe(sAttrs, 0) self.hStderr_r, hStderr_w = win32pipe.CreatePipe(sAttrs, 0) # set the info structure for the new process. StartupInfo = win32process.STARTUPINFO() StartupInfo.hStdInput = hStdin_r StartupInfo.hStdOutput = hStdout_w StartupInfo.hStdError = hStderr_w StartupInfo.dwFlags = win32process.STARTF_USESTDHANDLES # Mark doesn't support wShowWindow yet. # StartupInfo.dwFlags = StartupInfo.dwFlags | win32process.STARTF_USESHOWWINDOW # StartupInfo.wShowWindow = win32con.SW_HIDE # Create new output read handles and the input write handle. Set # the inheritance properties to FALSE. Otherwise, the child inherits # the these handles; resulting in non-closeable handles to the pipes # being created. pid = win32api.GetCurrentProcess() tmp = win32api.DuplicateHandle( pid, self.hStdin_w, pid, 0, 0, # non-inheritable!! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStdin_w) self.hStdin_w = tmp tmp = win32api.DuplicateHandle( pid, self.hStdout_r, pid, 0, 0, # non-inheritable! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStdout_r) self.hStdout_r = tmp tmp = win32api.DuplicateHandle( pid, self.hStderr_r, pid, 0, 0, # non-inheritable! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStderr_r) self.hStderr_r = tmp # start the process. print "creating process" cmdline = "%s %s" % (command, string.join(args, ' ')) hProcess, hThread, dwPid, dwTid = win32process.CreateProcess( None, # program cmdline,# command line None, # process security attributes None, # thread attributes 1, # inherit handles, or USESTDHANDLES won't work. # creation flags. Don't access the console. 0, # Don't need anything here. # If you're in a GUI app, you should use # CREATE_NEW_CONSOLE here, or any subprocesses # might fall victim to the problem described in: # KB article: Q156755, cmd.exe requires # an NT console in order to perform redirection.. environment, # new environment path, # new directory StartupInfo) # normally, we would save the pid etc. here... print "process created" # Child is launched. Close the parents copy of those pipe handles # that only the child should have open. # You need to make sure that no handles to the write end of the # output pipe are maintained in this process or else the pipe will # not close when the child process exits and the ReadFile will hang. win32file.CloseHandle(hStderr_w) win32file.CloseHandle(hStdout_w) win32file.CloseHandle(hStdin_r) self.outQueue = Queue.Queue() self.closed = 0 self.stdoutClosed = 0 self.stderrClosed = 0 threading.Thread(target=self.doWrite).start() win32.addEvent(self.hStdout_r, self, self.doReadOut) win32.addEvent(self.hStderr_r, self, self.doReadErr) def write(self, data): """Write data to the process' stdin.""" self.outQueue.put(data) def closeStdin(self): """Close the process' stdin.""" self.outQueue.put(None) def connectionLost(self): """Will be called twice, by the stdout and stderr threads.""" if not self.closed: win32.removeEvent(self.hStdout_r) win32.removeEvent(self.hStderr_r) abstract.FileDescriptor.connectionLost(self) print "connection lost" self.closed = 1 self.closeStdin() win32file.CloseHandle(self.hStdout_r) win32file.CloseHandle(self.hStderr_r) def doWrite(self): """Runs in thread.""" while 1: data = self.outQueue.get() if data == None: break try: win32file.WriteFile(self.hStdin_w, data, None) except win32api.error: break win32file.CloseHandle(self.hStdin_w) def doReadOut(self): """Runs in thread.""" try: hr, data = win32file.ReadFile(self.hStdout_r, 8192, None) except win32api.error: self.stdoutClosed = 1 if self.stderrClosed: return main.CONNECTION_LOST else: return self.handleChunk(data) def doReadErr(self): """Runs in thread.""" try: hr, data = win32file.ReadFile(self.hStderr_r, 8192, None) except win32api.error: self.stderrClosed = 1 if self.stdoutClosed: return main.CONNECTION_LOST else: return self.handleError(data) if __name__ == '__main__': win32.install() def printer(x): print "Got", repr(x) exe = win32api.GetModuleFileName(0) print exe p = Process(exe, ['-u', 'processtest.py'], None, None) print "ok, made process object" p.handleChunk = printer p.handleError = printer p.write("hello, world") p.closeStdin() main.run()

1 0

[Twisted-Python] twisted.internet.process for Win32
by Itamar Shtull-Trauring 10 Mar '02

10 Mar '02

I've attached my current code, which seems to be working. This will hopefully be checked in today, but some testing by others would be useful. """ Win32 process support. """ # Twisted imports from twisted.internet import task, abstract # System imports import win32api import win32pipe import win32file import win32process import win32security import win32con import win32event import pywintypes import msvcrt import os import sys import threading import Queue import string class Process(abstract.FileDescriptor): """A process that integrates with the Twisted event loop.""" buffer = '' def __init__(self, command, args, environment, path): # security attributes for pipes sAttrs = win32security.SECURITY_ATTRIBUTES() sAttrs.bInheritHandle = 1 # create pipes hStdin_r, self.hStdin_w = win32pipe.CreatePipe(sAttrs, 0) self.hStdout_r, hStdout_w = win32pipe.CreatePipe(sAttrs, 0) self.hStderr_r, hStderr_w = win32pipe.CreatePipe(sAttrs, 0) # set the info structure for the new process. StartupInfo = win32process.STARTUPINFO() StartupInfo.hStdInput = hStdin_r StartupInfo.hStdOutput = hStdout_w StartupInfo.hStdError = hStderr_w StartupInfo.dwFlags = win32process.STARTF_USESTDHANDLES # Mark doesn't support wShowWindow yet. # StartupInfo.dwFlags = StartupInfo.dwFlags | win32process.STARTF_USESHOWWINDOW # StartupInfo.wShowWindow = win32con.SW_HIDE # Create new output read handles and the input write handle. Set # the inheritance properties to FALSE. Otherwise, the child inherits # the these handles; resulting in non-closeable handles to the pipes # being created. pid = win32api.GetCurrentProcess() tmp = win32api.DuplicateHandle( pid, self.hStdin_w, pid, 0, 0, # non-inheritable!! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStdin_w) self.hStdin_w = tmp tmp = win32api.DuplicateHandle( pid, self.hStdout_r, pid, 0, 0, # non-inheritable! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStdout_r) self.hStdout_r = tmp tmp = win32api.DuplicateHandle( pid, self.hStderr_r, pid, 0, 0, # non-inheritable! win32con.DUPLICATE_SAME_ACCESS) # Close the inhertible version of the handle win32file.CloseHandle(self.hStderr_r) self.hStderr_r = tmp # start the process. print "creating process" cmdline = "%s %s" % (command, string.join(args, ' ')) hProcess, hThread, dwPid, dwTid = win32process.CreateProcess( None, # program cmdline,# command line None, # process security attributes None, # thread attributes 1, # inherit handles, or USESTDHANDLES won't work. # creation flags. Don't access the console. 0, # Don't need anything here. # If you're in a GUI app, you should use # CREATE_NEW_CONSOLE here, or any subprocesses # might fall victim to the problem described in: # KB article: Q156755, cmd.exe requires # an NT console in order to perform redirection.. environment, # new environment path, # new directory StartupInfo) # normally, we would save the pid etc. here... print "process created" # Child is launched. Close the parents copy of those pipe handles # that only the child should have open. # You need to make sure that no handles to the write end of the # output pipe are maintained in this process or else the pipe will # not close when the child process exits and the ReadFile will hang. win32file.CloseHandle(hStderr_w) win32file.CloseHandle(hStdout_w) win32file.CloseHandle(hStdin_r) self.outQueue = Queue.Queue() self.closed = 0 threading.Thread(target=self.doWrite).start() threading.Thread(target=self.doReadOut).start() threading.Thread(target=self.doReadErr).start() def write(self, data): """Write data to the process' stdin.""" self.outQueue.put(data) def closeStdin(self): """Close the process' stdin.""" self.outQueue.put(None) def connectionLost(self): """Will be called twice, by the stdout and stderr threads.""" if not self.closed: abstract.FileDescriptor.connectionLost(self) print "connection lost" self.closed = 1 self.closeStdin() win32file.CloseHandle(self.hStdout_r) win32file.CloseHandle(self.hStderr_r) def doWrite(self): """Runs in thread.""" while 1: data = self.outQueue.get() if data == None: break try: win32file.WriteFile(self.hStdin_w, data, None) except win32api.error: break win32file.CloseHandle(self.hStdin_w) def doReadOut(self): """Runs in thread.""" while 1: try: hr, data = win32file.ReadFile(self.hStdout_r, 8192, None) except win32api.error: task.schedule(self.connectionLost) return task.schedule(self.handleChunk, data) def doReadErr(self): """Runs in thread.""" while 1: try: hr, data = win32file.ReadFile(self.hStderr_r, 8192, None) except win32api.error: task.schedule(self.connectionLost) return task.schedule(self.handleError, data) if __name__ == '__main__': from twisted.internet import main def printer(x): print "Got", repr(x) exe = win32api.GetModuleFileName(0) print exe p = Process(exe, ['-u', 'processtest.py'], None, None) print "ok, made process object" p.handleChunk = printer p.handleError = printer p.write("hello, world") p.closeStdin() main.run()

2 5